Transcriptomic analysis supports the role of CATION EXCHANGER 1 in cellular homeostasis and oxidative stress limitation during cadmium stress

ABSTRACT

Investigation of genetic determinants of Cd tolerance in the Zn/Cd hyperaccumulator Arabidopsis halleri allowed the identification of the vacuolar Ca²⁺/H⁺ exchanger encoding CAX1 gene. CAX1 was proposed to interfere with the positive feedback loop between Reactive Oxygen Species (ROS) production and Cd-induced cytosolic Ca²⁺ spikes, especially at low external Ca²⁺ supply. In this study expression of genes involved in ROS homeostasis, cell wall composition, apoplastic pH regulation and Ca²⁺ homeostasis were monitored in Arabidopsis thaliana wild-type and cax1-1 knock-out mutant and in Arabidopsis halleri wild-type exposed to cadmium or in control conditions. Clustering the outputs of the expression analysis in a gene co-expression network revealed that CAX1 and genes involved in Ca²⁺ cellular homeostasis, apoplastic pH and oxidative stress response were highly correlated in A. thaliana, but not in A. halleri. Many of the studied genes were already highly expressed in A. halleri and/or their expression was not modified by exposure to Cd. The results further supported the role of CAX1 in the regulation of cytosolic ROS accumulation as well as the existence of different cell wall modifications strategies in response to Cd in Arabidopsis thaliana and halleri.

 

Genetic bases of Cd tolerance have been investigated in the Zn/Cd hyperaccumulator Arabidopsis halleri.¹ QTL analysis allowed the identification and the characterization of the heavy metal ATPase HMA4 and of the vacuolar Ca²⁺ /H⁺ exchanger encoding CAX1 4 as genetic determinants for Cd tolerance. In particular, CAX1 was proposed to interfere with the positive feedback loop shown between Reactive Oxygen Species (ROS) production and Cd-induced cytosolic Ca²⁺ spikes, especially at low external Ca²⁺ supply.⁴ CAX1 contribution to the cellular homeostasis, in particular as key-regulator of apoplastic Ca²⁺ and pH, was previously studied,^5,6 together with its capacity to shape cell wall extensibility and composition.

In this study expression analyses on A. thaliana WT (Col-0), A. thaliana CAX1 k.o. mutant (cax1-1) and A. halleri WT (Auby) were carried out to get a wider picture of Cd impact on steady state transcript level of genes possibly affected by CAX1 activity. The analysis was performed on root samples because (i) CAX1 expression was shown to be induced after Cd treatment only in this organ⁴ and (ii) Arabidopsis halleri and A. thaliana accumulated higher Cd concentration in roots than in shoots.⁷ A. halleri, A. thaliana WT and cax1-1 plants were grown in hydroponic solution containing 0.5 mM CaCl₂ and exposed to 10 µM CdSO₄ for 72 h (as in ⁴). A. halleri plants were also treated with 25 µM CdSO₄, given the higher tolerance of this species. Indeed, 10 µM and 25 µM CdSO₄ were previously reported to have similar physiological effects and toxicity on A. thaliana and A. halleri, respectively.⁸

Expression levels of genes involved in ROS homeostasis (RBOHs encoding NADPH Oxidase and CAT1 encoding Catalase 1), cell wall composition (PMEs encoding Pectin Methyl Esterase and CESA encoding Cellulose synthases), control of apoplastic pH (AHA1 and AHA2 encoding H⁺-ATPase 1 and 2) and Ca²⁺ homeostasis (CAX1; ACA8 encoding a Ca²⁺ -ATPase) were monitored through quantitative Real-Time PCR (qRT-PCR). List of primers used in this study and size of gene families are reported in Table S1A and B, respectively.

In order to compare A. thaliana WT, A. thaliana cax1-1 and A. halleri WT, a Principal Component Analysis (PCA)⁹ was performed on transcriptomic data (Fig. 1) with R programming language (Fig. 1). Samples were separated by the first 3 principal components (PC1, PC2 and PC3) accounting respectively for 41%, 24% and 18% of the variance (Fig. 1A). The PC1 clearly separated A. halleri samples from the others, while the PC2 pointed out the differences between control and Cd contaminated roots. The contribution of genes expression to samples distribution according to PC1 and PC2 is shown in Fig. 1B. It is worth noting that A. halleri samples clearly separated from A. thaliana (WT and cax1-1).

Figure 1.

Principal component analysis with qRT-PCR transcripts expression values carried out on A. thaliana, cax1-1 and A. halleri in control and 10 µM Cd contaminated conditions (3 biological replicates). (A). Samples (genotypes + conditions) 3D PCA scatterplot with the first 3 principal components. Percent of variance that can be explained is also reported for each component on the corresponding axes. (B). Bi-dimensional PCA plot. Samples distribution according to PC1 and PC2, and effect of each gene on samples position (indicated by eigenvectors, red arrows). Percentages designate the variance among the samples.

Interestingly, PCA distribution of cax1-1 samples upon Cd treatment seems to be deeply conditioned by transcript levels of genes involved in oxidative stress signaling (i.e., RBOHD and CAT1) and Ca²⁺ homeostasis (ACA8). These results well fit with the sensitive cax1-1 phenotype under Cd stress,⁴ which further supports the proposed role of CAX1 in the regulation of cytosolic ROS accumulation.

Expression profiles of genes that showed changes after Cd exposure or differential regulation among Arabidopsis genotypes are reported in Fig. 2.

Figure 2.

Expression analysis of genes involved in Ca homeostasis (A and B), apoplastic pH control (C and D), oxidative stress (E and F) and cell wall modifications (G and H) in control condition (gray bars), upon 10 µM CdSO₄ (white bars) and 25 µM CdSO₄ treatments (striped bars). Data are means ± SD (n = 3–6). Letters are for comparison among samples (Duncan test); * for difference among control samples and their correspondent Cd-treated samples (P < 0.05; t-test).

As shown in Baliardini et al. (2015)⁴ AtCAX1 expression was significantly induced in roots of A. thaliana by Cd treatment at low Ca²⁺ concentrations (0.5 mM), while AhCAX1 was already highly expressed in A. halleri in control conditions, and weakly induced by Cd (Fig. 2A). Different ACA8 expression levels were observed (Fig. 2B). ACA8 localizes in the plasma membrane and is responsible for Ca²⁺ efflux from the cytosol to the apoplastic space.^5,10 Exposure to Cd induced AtACA8 expression both in WT and cax1-1 most likely to limit the stress-induced increase of [Ca²⁺]_cytosol. Moreover, loss of CAX1 function impaired ACA8 expression in cax1-1, which was higher than in A. thaliana WT, both in control and Cd stress conditions. Interestingly, no change in AhACA8 transcript level was detected in A. halleri exposed to 10 µM or 25 µM CdSO₄. Considering the role of CAX1 in the maintenance of the apoplastic pH,^5,6 expression of AHA1 and AHA2 encoding 2 major proton pumps at the plasma membrane¹¹ was monitored. Cd exposure led to a significant increase of AHA2 transcript in A. thaliana and A. halleri, while AHA1 was induced by Cd only in A. halleri (Fig. 2C and D). On the other hand, AHAs were down-regulated in cax1-1. Cd-induced increase of AHAs transcripts in A. halleri and A. thaliana may be linked to the proton release into the cytosol in exchange of Ca²⁺ due to CAX1 activity.⁶

Expression of genes involved in oxidative stress response (i.e., RBOHs and CAT1) was also investigated. Cd contamination enhanced RBOHD expression in A. thaliana, as previously observed by Cuypers et al. (2011),¹² and not in A. halleri. cax1-1 roots highly expressed RBOHD already in control condition and exhibited the strongest RBOHD expression level upon Cd exposition, which correlated with the stronger ROS accumulation reported in Baliardini et al. (2015).⁴ Similarly, the expression of CAT1 increased more than 2-fold upon Cd treatment compared to control in cax1-1 while it did not show change in A. halleri.

Recent works revealed the importance of the cell wall compartment in metal tolerance and hyperaccumulation in A. halleri.^13-15 Among the tested genes involved in cell wall modifications, CESA4 and PME1 transcripts (Fig. 2E and F) were significantly downregulated in A. thaliana genotypes (both WT and cax1-1) exposed to Cd, while they were induced in A. halleri. These results suggest that A. thaliana and A. halleri may activate contrasting strategies involving cell wall modifications in response to Cd, as already suggested by Meyer et al. (2015).¹⁴ Likely, highly rated proton and Ca²⁺ efflux mediated by AHAs and ACA8 may affect cell wall composition and PMEs activity.^5,18-20 PMEs, which are responsible for the removal of methyl groups from pectins, may allow the exposure of free charges, usually target of Ca²⁺ bounds.^20-22 It is conceivable that the general increase of PMEs and CESAs transcripts expression in response to Cd treatment in A. halleri may facilitate Cd binding by displacing Ca²⁺ ions, as suggested in Isaure et al. (2015).¹⁵ As secondary consequence, higher PMEs and CESAs activity may cause the stiffening of the cell wall in order to build up a barrier to metal uptake into the cytosol.^{14,15, 23,24} Further tests such as enzyme activity assays, microscope and ionomics analyses and reverse genetics approaches may help to uphold these speculations.

Clustering the outputs of the expression analysis in a gene co-expression network (Fig. 3) revealed that CAX1 and genes involved in Ca²⁺ cellular homeostasis, apoplastic pH and oxidative stress response were highly correlated in A. thaliana (Fig. 3A), but not in A. halleri (Fig. 3B). Indeed, while in A. thaliana CAX1, CML41, ACA8, AHA1, AHA2, PME41 and RBOHD transcripts were induced in response to Cd, they were all constitutively (in the absence of Cd exposure) highly expressed or showed unchanged transcripts levels after Cd contamination in A. halleri. This evidence supports the presence of an altered basal cellular homeostasis in A. halleri compared to A. thaliana already in control condition, which probably allows a more efficient response to Cd exposure as suggested by other studies.^15,16 For A. halleri, the correlation network analysis did not show significant differences upon Cd treatment at 10 µM or 25 µM, (Fig. 3B and Fig. 3C respectively). Indeed, despite the higher number of correlated transcripts found at 25 µM Cd (Fig. 3C), the number of genes whose expression was higher in Cd-treated samples compared to the untreated, was substantially unchanged compared to what observed at 10 µM.

Figure 3.

Co-expression network of Arabidopsis thaliana and A. halleri genes expression. Pearson correlations and adjusted p values were calculated for each gene pairs based on their qRT-PCR expression (3 genotypes, 2 conditions, 3 biological replicates). False discovery rate (FDR) <0.05; Pearson correlation coefficient r > |0.9|. Continuous lines indicate positive correlation between genes, while dashed lines indicate a negative correlation. Red and blue boxes indicate genes more expressed in Cd contaminated and control samples, respectively (log₂ Cd/Ctr).

On the whole, the expression results suggest that several mechanisms affected by Cd treatment (Ca homeostasis, redox balance control and modifications at the level of cell wall and apoplastic pH), are also influenced by CAX1 activity in Arabidopsis. Furthermore Cd induced more changes in the expression of genes analyzed in this study in the sensitive A. thaliana rather than in the tolerant A. halleri. These evidences further support that A. halleri tolerance mechanisms are not likely to be “switched on/off” depending on Cd exposure, but are rather already present in the absence of metal excess.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Funding

The authors thank the Fonds national de la Recherche scientifique for financial support (PDR T.0206.13 and fellowships of C. B. and M.C)